Intelligent filling method and system for data missing based on multiple interpolation fusion

Abstract

The application belongs to the technical field of data processing, and provides an intelligent filling method and system for data missing based on multiple interpolation fusion. The method comprises the following steps: acquiring all data sequences to be interpolated and performing distribution statistics, and dividing the distribution statistics results into a missing group and a continuous group; learning the missing distribution characteristics in the missing group, and then constructing a bionic missing group based on the missing distribution characteristics and the continuous group, while recording the data group of the missing position as an answer set; calling each filling method in a preset filling method set to process the bionic missing group, obtaining an interpolation data set, comparing the interpolation data set with the data in the answer set, determining whether each filling value in the interpolation data set is correct, and then calculating a group of accuracy rates of each sequence in the bionic missing group; screening out the interpolation method with the highest accuracy rate of each sequence in the bionic missing group for corresponding interpolation processing, and finally obtaining an accurate filling set.

Description

Technical Field

[0001] This invention belongs to the field of data processing technology, and in particular relates to an intelligent method and system for filling missing data based on multiple interpolation fusion. Background Technology

[0002] The statements in this section are merely background information related to the present invention and do not necessarily constitute prior art.

[0003] In routine data cleaning and processing, a sequence of null values ​​frequently occurs. This may be due to issues during data cleaning, omissions during initial data collection, or accidental deletion. Regardless of the cause, null value sequences lead to unreliable research results. Missing data causes consecutive datasets to lose much useful information, resulting in significant discrepancies between the analyzed data and the true values. Therefore, a method to address these problems is urgently needed.

[0004] There are many interpolation methods available, such as linear interpolation, mean interpolation, mode interpolation, and cubic spline interpolation. Each method has its own application conditions. For example, mean interpolation is suitable for data sequences with small numerical variations, while linear interpolation is suitable for data sequences with a linear distribution. However, in practical applications, data is complex and variable, consisting of a collection of highly complex conditions. It is impossible to fit all missing sequences using a single method. The function changes at different times, which can lead to significant biases in the results if only one interpolation method is used to fill in the data.

[0005] With the rise of artificial intelligence, deep learning networks are also used to predict corresponding results. For example, LSTM networks are commonly used to predict time series and can sometimes be used for missing data imputation tasks. However, these complex neural networks often require massive amounts of high-quality data and are sometimes not suitable for small sample data, and their accuracy is sometimes not high. Summary of the Invention

[0006] To address the technical problems mentioned above, this invention provides an intelligent data filling method and system based on multiple interpolation fusion, which can improve the accuracy of data filling.

[0007] To achieve the above objectives, the present invention adopts the following technical solution:

[0008] The first aspect of the present invention provides an intelligent filling method for missing data based on multiple interpolation fusion.

[0009] A smart imputation method for missing data based on multiple interpolation fusion, comprising:

[0010] Obtain all data sequences to be interpolated and perform distribution statistics, then divide the distribution statistics results into missing groups and continuous groups; wherein, the data sequences to be interpolated contain independent and identically distributed complete data sequence samples;

[0011] Learn the missing distribution characteristics in the missing groups, and then construct a biomimetic missing group based on the missing distribution characteristics and continuous groups. At the same time, record the data group of missing positions and use it as the answer set.

[0012] The system retrieves various filling methods from the preset filling method set to process the biomimetic missing group, obtains the interpolated data set, and compares it with the data in the answer set to determine whether each filling value in the interpolated data set is correct. Then, it calculates the accuracy of a set of sequences in the biomimetic missing group.

[0013] The interpolation method with the highest accuracy for each sequence in the biomimetic missing group is selected and interpolated accordingly to obtain the accurate filled set.

[0014] As one implementation method, the missing data distribution characteristics in missing groups are learned through a biomimetic algorithm. The process is as follows:

[0015] Filter out the shortest sequence in the missing group, and truncate all data in the missing group to the same length as the shortest sequence by left alignment;

[0016] Then record the index of the missing location, which is the learned missing distribution feature.

[0017] As one implementation method, the process of constructing a biomimetic missing group is as follows:

[0018] First, process the sequences in the continuous group into sequences of the same length as the shortest sequence in the missing group;

[0019] The missing distribution index is then applied to the equal-length sequences of continuous groups to obtain the biomimetic missing groups.

[0020] As one implementation method, the principle for determining the correctness of each filler value in the interpolated dataset is as follows:

[0021] The error between the interpolated dataset and the data in the answer set is within a preset range.

[0022] As one implementation method, the process of selecting the interpolation method with the highest accuracy for each sequence in the biomimetic missing group is as follows:

[0023] The accuracy of each sequence in the biomimetic missing group constitutes an accuracy set;

[0024] The sequence of accuracy sets is sorted according to the accuracy, and the interpolation method with the highest accuracy is selected for each subset.

[0025] As one implementation method, accuracy = number of correct fills / (number of correct fills + number of incorrect fills).

[0026] A second aspect of the present invention provides an intelligent data filling system based on multiple interpolation fusion.

[0027] An intelligent data missing filling system based on multiple interpolation fusion includes:

[0028] The distribution statistics module is used to acquire all data sequences to be interpolated and perform distribution statistics, and divide the distribution statistics results into missing groups and continuous groups; wherein, the data sequences to be interpolated contain independent and identically distributed complete data sequence samples;

[0029] The biomimetic learning module is used to learn the missing distribution characteristics in the missing groups, and then construct the biomimetic missing groups based on the missing distribution characteristics and continuous groups, while recording the data groups of missing positions as the answer set;

[0030] The filling comparison module is used to retrieve various filling methods from the preset filling method set to process the bionic missing group, obtain the interpolated data set, and compare it with the data in the answer set to determine whether each filling value in the interpolated data set is correct, and then calculate the accuracy of a set of sequences in the bionic missing group.

[0031] The interpolation filtering module is used to select the interpolation method with the highest accuracy for each sequence in the biomimetic missing group, perform corresponding interpolation processing, and finally obtain the accurate filled set.

[0032] A third aspect of the present invention provides a computer-readable storage medium.

[0033] A computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the intelligent imputation method for data missing data based on multiple interpolation fusion as described above.

[0034] A fourth aspect of the present invention provides a computer device.

[0035] A computer device includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the steps of the intelligent data missing filling method based on multiple interpolation fusion as described above.

[0036] Compared with the prior art, the beneficial effects of the present invention are:

[0037] This invention addresses the issue of missing data by proposing a method that intelligently analyzes a given data sequence, dividing the data into two sets: missing groups and continuous groups. The missing groups are then statistically analyzed for their missing distribution. This statistical distribution is applied to the continuous groups to create artificially constructed missing groups. Simultaneously, the true values ​​at the corresponding missing positions are recorded as the answer set. Various known interpolation methods are then used to interpolate the artificially constructed missing groups, and all interpolation results are compared with the true values ​​to calculate accuracy. The algorithm then identifies the interpolation method with the highest accuracy for different missing distributions and assigns an optimal interpolation method to each missing distribution type in the current data. This method is then applied to the initially created missing groups, using the corresponding interpolation method for each missing distribution to achieve the best interpolation effect. This scheme eliminates the need for data labeling, autonomously learning the characteristics of missing distributions and matching the optimal interpolation method to the missing data, thus improving the filling effect.

[0038] Advantages of additional aspects of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. Attached Figure Description

[0039] The accompanying drawings, which form part of this invention, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an improper limitation of the invention.

[0040] Figure 1 This is a specific distribution diagram of the missing data in an embodiment of the present invention;

[0041] Figure 2 This is an accuracy distribution diagram of various filling methods according to embodiments of the present invention;

[0042] Figure 3 This is a flowchart of the intelligent data filling method based on multiple interpolation fusion according to an embodiment of the present invention;

[0043] Figure 4 This is a schematic diagram of the intelligent data filling principle based on multiple interpolation fusion in an embodiment of the present invention;

[0044] Figure 5 This is a schematic diagram of the intelligent data filling system based on multiple interpolation fusion according to an embodiment of the present invention. Detailed Implementation

[0045] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0046] It should be noted that the following detailed description is illustrative and intended to provide further explanation of the invention. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0047] It should be noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the scope of exemplary embodiments according to the invention. As used herein, the singular form is intended to include the plural form as well, unless the context clearly indicates otherwise. Furthermore, it should be understood that when the terms "comprising" and / or "including" are used in this specification, they indicate the presence of features, steps, operations, devices, components, and / or combinations thereof.

[0048] Example 1

[0049] Combination Figure 3 and Figure 4 This embodiment provides an intelligent data filling method based on multiple interpolation fusion, which specifically includes the following steps:

[0050] Step 1: Obtain all data sequences to be interpolated and perform distribution statistics, and divide the distribution statistics results into missing groups and continuous groups; wherein, the data sequences to be interpolated contain independent and identically distributed complete data sequence samples.

[0051] This invention is applicable to data missing problems in various scenarios. This article uses intraoperative blood pressure data as an example, but is not limited to blood pressure data. This article selects intraoperative blood pressure data from 539 patients undergoing craniotomy in three hospitals. The blood pressure data was recorded every 10 seconds, that is, 6 data points were collected per minute. Among them, the data of 165 patients were continuous and without missing data, while the data of the remaining 374 patients had varying degrees of missing data.

[0052] For example: Given intraoperative blood pressure data of multiple patients, which is a data sequence S to be interpolated, it needs to contain some complete data samples S. b Furthermore, the blood pressure readings must satisfy the independent and identically distributed condition with the sequence S to be interpolated, meaning that these blood pressure readings must be collected from the same device.

[0053] Load all data sequences S1, S2, ..., S to be interpolated n These sequences are then processed using a missing sequence distribution statistical algorithm. Here, the missing sequences are categorized into 0 minutes (i.e., continuous data without missing sequences), 1 minute, 2 minutes, ..., 30 minutes, and the results are denoted as d0, d1, d2, ..., d... 30Here, 1 minute refers to a sequence with 1 to 6 missing data entries within 1 minute. Similarly, 30 minutes means a sequence with a cumulative total of approximately 1 to 180 missing data entries within 30 minutes. Since typical surgeries last 3 to 4 hours, the anesthesiologist will monitor the patient's blood pressure in real time during the procedure.

[0054] In practical applications, a missing time span of 30 minutes is relatively rare. Specifically, in this embodiment, the maximum missing time is set to 30 minutes. However, there is no upper limit to this method.

[0055] The distribution statistics results are divided into missing groups and continuous groups, where the missing group S a ={a1, a2, ..., a 30}, continuous sets S without missing elements b ={b0}.

[0056] Step 2: Learn the missing distribution characteristics in the missing groups, and then construct a biomimetic missing group based on the missing distribution characteristics and continuous groups. At the same time, record the data group of missing positions as the answer set.

[0057] In some specific implementation processes, in step 2, the missing distribution characteristics in the missing groups are learned through a biomimetic algorithm. The process is as follows:

[0058] Filter out the shortest sequence in the missing group, and truncate all data in the missing group to the same length as the shortest sequence by left alignment;

[0059] Then record the index of the missing location, which is the learned missing distribution feature.

[0060] The biomimetic algorithm specifically refers to: targeting S a The missing data distribution is shown in the figure, and S is taken as the missing data distribution. a The shortest sequence a in the sequence s , will S a All data is left-aligned and cropped using the a-axis. s A sequence of equal length can then be used to record the index S of the missing positions. p ={p1,p2,…,p 30 These sequences are the learned missing distribution features.

[0061] Specifically, the process of constructing the biomimetic missing group is as follows:

[0062] First, process the sequences in the continuous group into sequences of the same length as the shortest sequence in the missing group;

[0063] The missing distribution index is then applied to the equal-length sequences of continuous groups to obtain the biomimetic missing groups.

[0064] The missing distribution information S is obtained through the aforementioned biomimetic algorithm.p Application to S b In terms of data, specifically S... b The data in the set is obtained according to the S obtained in the previous step. p The process is processed to make S b The sequence is processed into a s Equal-length sequence S bs This allows the missing distribution index S to be displayed. p Applied to set S bs And obtained the artificially constructed biomimetic missing group S c ={c1,c2,…,c 30}, and simultaneously record the data set Q = {q1, q2, ..., q} at the missing positions. 30 Let} be denoted as the answer set Q, where the subscripts 1 to 30 have the same meaning as the subscripts mentioned earlier.

[0065] It is understandable here that, in addition to the biomimetic algorithm mentioned above, other algorithms can also be used to learn the missing distribution characteristics in missing groups and to construct biomimetic missing groups, which will not be elaborated here.

[0066] Step 3: Retrieve each filling method from the preset filling method set to process the biomimetic missing group, obtain the interpolated data set, and compare it with the data in the answer set to determine whether each filling value in the interpolated data set is correct, and then calculate the accuracy of a set of sequences in the biomimetic missing group.

[0067] The set of filling methods includes, but is not limited to, eight filling interpolation algorithms: mean filling, mode filling, median filling, previous filling, next filling, linear interpolation, quadratic B-spline curve filling, and cubic B-spline curve filling. These are denoted as F. mean F mode F median F previous F next F linear F quadratic F cubic , thus obtaining the set of filling methods F = {F mean ,F mode ,F median ,F previous ,F next ,F linear ,F quadratic ,F cubic}

[0068] Accuracy rate = number of correct fills / (number of correct fills + number of incorrect fills).

[0069] Furthermore, the principle for determining the correctness of each filler value in the interpolated dataset is as follows:

[0070] The error between the interpolated dataset and the data in the answer set is within a preset range.

[0071] It should be noted that the preset range varies depending on the nature of the data, and should be set by those skilled in the art.

[0072] Step 4: Select the interpolation method with the highest accuracy for each sequence in the biomimetic missing group, perform corresponding interpolation processing, and finally obtain the accurate filled set.

[0073] In step 4, the process of selecting the interpolation method with the highest accuracy for each sequence in the biomimetic missing group is as follows:

[0074] The accuracy of each sequence in the biomimetic missing group constitutes an accuracy set;

[0075] The sequence of accuracy sets is sorted according to the accuracy, and the interpolation method with the highest accuracy is selected for each subset.

[0076] Regarding the biomimetic missing group S mentioned above c The data are respectively applied to the filling interpolation algorithm selected in the previous step, and the interpolated data set P is compared with the answer set Q. The error is ±5mmHg (this value is the clinical error value provided by the hospital doctor). Data that meets the following formula (1) is considered to be filled correctly, otherwise it is considered to be filled incorrectly.

[0077] |GT-Pre|<5mmHg (1)

[0078] Where GT is the real data from the answer set Q, and Pre is the biomimetic data value filled by the algorithm from the set P.

[0079] Therefore, in S respectively c For each sequence in the sequence, an accuracy rate can be obtained, where accuracy = number of correctly filled items / (number of correctly filled items + number of incorrectly filled items), denoted by S. c Taking c1 as an example, we can obtain N. C1 ={n 1mean ,n 1mode ,n 1median ,n 1previous ,n 1next ,n 1linear ,n 1quadratic ,n 1cubic},…,N C30 ={n 30mean ,n 30mode ,n 30median ,n 30previous ,n30next ,n 30linear ,n 30quadratic ,n 30cubic The accuracy set N = {N} is composed of} C1 N C2 ,…,N C30};

[0080] For the sequences of set N, sort them according to their accuracy, and for each subset, select the interpolation method with the highest accuracy, that is, for set S... c c1, c2, ..., c 30 Select the interpolation method F with the highest accuracy respectively. best ={F 1best ,F 2best ,…,F 30best};

[0081] Through the steps above, the algorithm autonomously learns the true data distribution of the data sequence S with missing sets, and matches the optimal set of interpolation methods F for various situations. Based on the information of the set of interpolation methods F, it can interpolate data in S with truly missing data. a By fusing data from dataset S and applying the optimal imputation method from F based on the missing data, this method achieves the best results compared to a single interpolation approach, ultimately yielding an accurately imputed set S. d S d ={d1,d2,…,d 30}

[0082] This study selected blood pressure data from 539 patients who underwent craniotomy at a tertiary hospital. Analysis of this data revealed the distribution of missing data, such as... Figure 1 As shown.

[0083] It was found that the data loss was mainly concentrated in the 1-10 minute range, accounting for over 90% of the total missing data. In fact, we solved 90% of the missing sequence problems by only addressing the missing data within this 10-minute range. Using ±5 mmHg as the error margin, we calculated the accuracy of using a single imputation method and the intelligent imputation method separately. Specific results are as follows... Figure 2 .

[0084] The results show that the effects of using a single filling method are inconsistent. The linear interpolation filling method is more effective among these single filling methods. Our intelligent filling method, on the other hand, fully leverages the advantages of each filling method through the intelligent analysis module, achieving a 1+1>2 effect and achieving the best accuracy among these methods.

[0085] It is worth noting that although this embodiment only selects 8 interpolation methods, it is not limited to these 8. Depending on the complexity of the dataset to be processed, more supporting means can be supported and more scenarios can be fitted according to the method provided in this article.

[0086] Example 2

[0087] like Figure 5 As shown, this embodiment provides an intelligent data filling system based on multiple interpolation fusion methods, which specifically includes the following modules:

[0088] The distribution statistics module is used to acquire all data sequences to be interpolated and perform distribution statistics, and divide the distribution statistics results into missing groups and continuous groups; wherein, the data sequences to be interpolated contain independent and identically distributed complete data sequence samples;

[0089] The biomimetic learning module is used to learn the missing distribution characteristics in the missing groups, and then construct the biomimetic missing groups based on the missing distribution characteristics and continuous groups, while recording the data groups of missing positions as the answer set;

[0090] The filling comparison module is used to retrieve various filling methods from the preset filling method set to process the bionic missing group, obtain the interpolated data set, and compare it with the data in the answer set to determine whether each filling value in the interpolated data set is correct, and then calculate the accuracy of a set of sequences in the bionic missing group.

[0091] The interpolation filtering module is used to select the interpolation method with the highest accuracy for each sequence in the biomimetic missing group, perform corresponding interpolation processing, and finally obtain the accurate filled set.

[0092] Specifically, in the biomimetic learning module, the missing data distribution characteristics in the missing data group are learned through a biomimetic algorithm, and the process is as follows:

[0093] Filter out the shortest sequence in the missing group, and truncate all data in the missing group to the same length as the shortest sequence by left alignment;

[0094] Then record the index of the missing location, which is the learned missing distribution feature.

[0095] It should be noted that each module in this embodiment corresponds one-to-one with each step in Embodiment 1, and their specific implementation processes are the same, so they will not be repeated here.

[0096] Example 3

[0097] This embodiment provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps in the intelligent data missing filling method based on multiple interpolation fusion as described in Embodiment 1 above.

[0098] Example 4

[0099] This embodiment provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, it implements the steps in the intelligent data missing filling method based on multiple interpolation fusion as described in Embodiment 1 above.

[0100] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of hardware embodiments, software embodiments, or embodiments combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage and optical storage) containing computer-usable program code.

[0101] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0102] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. The storage medium can be a magnetic disk, optical disk, read-only memory (ROM), or random access memory (RAM), etc.

[0103] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. An intelligent filling method for data missing based on multiple interpolation fusion, characterized in that, The method comprises the following steps: Obtain all intraoperative blood pressure data sequences to be interpolated and perform distribution statistics, and divide the distribution statistics into a missing group and a continuous group; wherein the intraoperative blood pressure data sequences to be interpolated contain independent and identically distributed complete intraoperative blood pressure data sequence samples; the complete intraoperative blood pressure data sequence samples and the intraoperative blood pressure data sequences to be interpolated are obtained by the same device; Learn the missing distribution characteristics in the missing group, and then construct a biomimetic missing group based on the missing distribution characteristics and the continuous group, and record the data group of the missing position as an answer set; Retrieve each filling method in the preset filling method set to process the biomimetic missing group, obtain an interpolation data set, and compare it with the intraoperative blood pressure data in the answer set to determine whether each filling value in the interpolation data set is correct, and then calculate a group of accuracy rates of each sequence in the biomimetic missing group; Select the interpolation method with the highest accuracy rate of each intraoperative blood pressure data sequence in the biomimetic missing group for corresponding interpolation processing, and finally obtain an accurate filling set; The longest missing time of the intraoperative blood pressure data to be interpolated is set to 30 minutes, and the blood pressure data is recorded every 10 seconds; Data satisfying formula (1) is considered to be filled correctly, otherwise it is considered to be filled incorrectly: ， Wherein GT is the true data from the answer set Q, and Pre is the biomimetic data value filled by the algorithm from the set P; The process of learning the missing distribution characteristics in the missing group by the biomimetic algorithm is as follows: Select the shortest sequence in the missing group, and cut all data in the missing group to the same length as the shortest sequence in a left alignment manner; Then record the missing position index, that is, the learned missing distribution characteristics; The process of constructing the biomimetic missing group is as follows: First, process the intraoperative blood pressure data in the continuous group into a sequence with the same length as the shortest sequence in the missing group; Then apply the missing distribution index to the equal-length sequence of the continuous group to obtain the biomimetic missing group; The principle for judging whether each filling value in the interpolation data set is correct is that the error between the interpolation data set and the data in the answer set is within a preset range; The process of selecting the interpolation method with the highest accuracy rate of each sequence in the biomimetic missing group is as follows: A group of accuracy rates of each sequence in the biomimetic missing group constitutes an accuracy rate set; The sequences in the accuracy rate set are sorted according to the accuracy rate, and the interpolation method with the highest accuracy rate is selected for each sub-set. Accuracy rate = number of correct filling / (number of correct filling + number of incorrect filling).

2. The method for intelligent filling of data missing based on multiple interpolation fusion of claim 1, wherein, The method comprises the following steps:

3. A system for intelligent filling of missing data based on multiple interpolation fusion, adopting the method for intelligent filling of missing data based on multiple interpolation fusion according to any one of claims 1-2, characterized in that, A distribution statistics module is used to obtain all intraoperative blood pressure data sequences to be interpolated and perform distribution statistics, and divide the distribution statistics into a missing group and a continuous group; wherein the intraoperative blood pressure data sequences to be interpolated contain independent and identically distributed complete intraoperative blood pressure data sequence samples; the complete intraoperative blood pressure data sequence samples and the intraoperative blood pressure data sequences to be interpolated are obtained by the same device; A biomimetic learning module is used to learn the missing distribution characteristics in the missing group, and then construct a biomimetic missing group based on the missing distribution characteristics and the continuous group, and record the data group of the missing position as an answer set; ​ The filling comparison module is configured to call each filling method in the preset filling method set to process the biomimetic missing group, obtain an interpolation data set, and compare the interpolation data set with intraoperative blood pressure data in the answer set to determine whether each filling value in the interpolation data set is correct, and further calculate a group of accuracy rates of each sequence in the biomimetic missing group; The interpolation screening module is configured to screen out an interpolation method with the highest accuracy rate for each sequence of intraoperative blood pressure data in the biomimetic missing group to perform corresponding interpolation processing, and finally obtain an accurate filling set.

4. The intelligent filling of missing data based on multiple interpolation fusion system as claimed in claim 3, wherein, In the biomimetic learning module, the missing distribution characteristics in the missing group are learned through a biomimetic algorithm, and the process is as follows: The shortest sequence in the missing group is screened out, and all data in the missing group are truncated to have the same length as the shortest sequence in a left alignment manner; Then, the position index of the missing data is recorded, i.e., the learned missing distribution characteristics.

5. A computer-readable storage medium having stored thereon a computer program, characterized in that, The program is executed by the processor to implement the steps in the intelligent filling method for data missing based on multiple interpolation fusion according to any one of claims 1-2.

6. A computer device, comprising: The computer program is stored in the memory and executable on the processor, and the processor implements the steps in the intelligent filling method for data missing based on multiple interpolation fusion according to any one of claims 1-2 when executing the program.

Images